[03/20/2024-06:00:35] [W] [TRT] parsers/onnx/onnx2trt_utils.cpp:375: Your ONNX model has been generated with INT64 weights, while TensorRT does not natively support INT64. Attempting to cast down to INT32.
[03/20/2024-06:00:35] [I] Finish parsing network model
[03/20/2024-06:00:38] [I] FP32 and INT8 precisions have been specified - more performance might be enabled by additionally specifying --fp16 or --best
[03/20/2024-06:00:41] [W] [TRT] Calibrator won't be used in explicit precision mode. Use quantization aware training to generate network with Quantize/Dequantize nodes.
[03/20/2024-17:33:14] [I] [TRT] [MemUsageChange] Init cuBLAS/cuBLASLt: CPU +6, GPU -21, now: CPU 1416, GPU 25098 (MiB)
[03/20/2024-17:33:19] [I] [TRT] [MemUsageChange] Init cuDNN: CPU +2, GPU +26, now: CPU 1418, GPU 25124 (MiB)
[03/20/2024-17:33:20] [I] [TRT] Local timing cache in use. Profiling results in this builder pass will not be stored.
[03/20/2024-18:14:25] [I] [TRT] Total Host Persistent Memory: 372224
[03/20/2024-18:14:25] [I] [TRT] Total Device Persistent Memory: 0
[03/20/2024-18:14:25] [I] [TRT] Total Scratch Memory: 0
[03/20/2024-18:14:25] [I] [TRT] [MemUsageStats] Peak memory usage of TRT CPU/GPU memory allocators: CPU 418 MiB, GPU 216 MiB
[03/20/2024-18:14:26] [I] [TRT] [BlockAssignment] Algorithm ShiftNTopDown took 45.2042ms to assign 7 blocks to 184 nodes requiring 34969600 bytes.
[03/20/2024-18:14:26] [I] [TRT] Total Activation Memory: 34969600
[03/20/2024-18:14:28] [W] [TRT] TensorRT encountered issues when converting weights between types and that could affect accuracy.
[03/20/2024-18:14:28] [W] [TRT] If this is not the desired behavior, please modify the weights or retrain with regularization to adjust the magnitude of the weights.
[03/20/2024-18:14:28] [W] [TRT] Check verbose logs for the list of affected weights.
[03/20/2024-18:14:28] [W] [TRT] - 25 weights are affected by this issue: Detected values which are outside of int8_t range and clipped them to int8_t range.
[03/20/2024-18:14:28] [I] [TRT] [MemUsageChange] TensorRT-managed allocation in building engine: CPU +83, GPU +84, now: CPU 83, GPU 84 (MiB)
[03/20/2024-18:20:20] [I] Engine built in 44431.5 sec.